 Welcome to Precision Agriculture in the Southeast. I'm here with Dr. John Fulton, our precision ag specialist for Auburn University and our Extension leader. And our topic today is yield monitoring, John. But before we get into that, we were talking before it started guidance system. Why this dome? You know, I got guidance system in my phone here, and it's a little bitty. Why the big, what's under the dome? Well, there's several components that's all integrated in here, number one, you can see deer's radio, they use the the transmit between the base station and the rover, which would be on the tractor, but internally here to the the donors, not only the GPS antenna, there's receiver and also built in some basically mechanical features to minimize any of the errors that are associated when we think about the transmission of electromagnetic waves or the bouncy of those waves off of metal. We want to try and do everything that we can. So it's direct signal, not a reflected signal. We want the direct signal being caught by that antenna and being used by the receiver to compute that and anything that can be done to to improve that positioning, as we talked about earlier, is going to keep us down to that centimeter level accuracy, especially when we get to at that. It's not something round under there. And there's like a metal plate right there. They probably have some type of metal plate or plane in there that helps eliminate some of those potential errors. Well, sorry for the distraction, John. Tell us about yield monitoring. Well, yield monitoring is our lesson here. And, you know, really, it's it's a key ingredient, key tool in any precision egg program at the farm. We've really seen the growth of yield monitoring in the last few years. And a lot of that stem from number one, folks really building their precision egg programs at their farm and the requirement to have yield monitor data in order to do things like how do I develop zones? How do I go out and apply inputs on a variable rate aspect? So yield monitoring really provides a key ingredient to help direct and develop those strategies. So we're going to kind of go through the parts of this from a yield monitoring, just what are the components? That's an important thing. And hopefully along the way, tell a little bit about some of the potential errors or what to look out for. Because at the end of the day, we want to have good quality data that can be used back in the program or within the business to make decisions with. So I think it's imperative right off the front that while everyone has this concept of what yield monitoring and we're going to talk about yield mapping at another module. But, you know, it's very important to understand what the basic components, what's measuring various parameters on the harvester to come up with this yield estimate, whether that's bushels per acre in grain or pounds per acre in cotton. There's particular sensors that are used or in most cases, multiple sensors used to derive that yield. But to understand those potential errors, because at the end of the day, if we're going to use this back in our program to make decisions or help us push forward in our precision ag program, you know, it's only as good as the data that's provided. So we want to make sure that in this lesson that we kind of hit those one and two, what those components are, what those potential errors are, just to be in recognition of that. So as I'm out harvesting, maybe I need to check something in the morning or every morning to ensure that I'm getting quality data. In the business today, you know, as we stand, the big push is quality data. And really coming at this because so many of these companies are providing services today and are really requiring farmers to provide quality yield data. So this, you know, right off the bat, we talk about what a yield map, it really can, you know, it's a multitude of things that we can start to drive. Essentially it's the ability to look at yield spatially on a field by field basis. But the interesting point that we got to kind of keep in mind, it does not point out what the yield limiting factor is. And so it only shows you what the response is. And so ground truth either going back out and looking at those areas, maybe low yielding or high yielding and kind of looking at those in a separate lens today versus just treating that field as a whole is important. But the yield map is not going to tell you what the due cause of that is. So yield monitoring, you know, I think it's just critical today as we see all these services that are rolled out by companies that they all require yield data, you know, that piece. It's very important. If you, you know, we try and use a lot of other data layers and absence of yield monitoring. But at the end of the day, the yield map shows the money. Where the money's coming from as we've been talking about market. If we can figure out how to either generate more money or not invest as much, you know, today's business operation at the farm is how do I best invest my money to get the maximum return? And that's going to be critical here. We basically collect in geo reference going back to the DGPS. We're using DGPS. We got a point. We're attaching some yield estimate to that point. And so all of a sudden now we can generate a map and do some spatial analysis and understanding the variability out in the field. We provide instantaneous yield and moisture measurements. In the case for grain, for cotton yield monitors, we're just getting yield. That pounds per second or pounds per acre. John compared to, you know, just the farmer driving a combine without a yield monitor. He knows when the combine, he can hear the grain going through and he knows when he's hit a sweet spot versus he's running a long time. But this really helps a farmer understand his farm. He knows where his sweet spots are and where his trouble spots are. And you take wet years and dry years and over time you learn how, what to expect, how to treat that land, what to do to improve it. So it's really not just an issue. It's not just data. Now, one of my favorite things is to look on Facebook and like in wheat harvest and somebody's taking a picture of their yield monitor and it's showing 120 bushels. That's like, ah, yeah. I bet that's not farm wide. That's a sweet spot. Absolutely. But it really helps people know what they're doing. Well, and I think it's an instant feedback. Absolutely. That instantaneous feedback in the cab. And, you know, probably the most common point the farmers have made is I never knew that much difference existed. I knew there was highs and lows, but, you know, there was 50 or 70 bushel from a high to low perspective out in that field. And so it really brings the quantification of that into sight. And wow, hey, you know, maybe I need to, can I capitalize knowing that and change some management strategies. You know, at the end of the day, we want that yield map. So that's kind of yield monitor. We're just trying to get spatial context on yield across the field, bringing that back and using that as a data layer today and decision making. Key uses, and I'm just going to throw out three. Really, it's, you know, you'll find out that once you got the yield monitor, there's a lot of uses for it, whether that's loading trucks out, knowing what the real moisture content of the grain is. But from a kind of a strategic standpoint, diagnosing crop production problems, you know, had a weed issue here. I got a fertility issue. I got a pH issue, but identifying those and correcting those and again managing to what the crop's telling us or what at the end of the year that we're seeing. What's that spatial yield potential? You know, ultimately, we talk about nutrient management. You know, can we take some of this data that we collect over time and really start to establish our nitrogen management? Maybe even potentially P&K, but more importantly, making sure our pH is maintaining within range across the whole field. The other thing is ultimately is establishing what your true yield potential is for the field. And just think about that. If I can start to drive that and have a lot better history and knowledge for that field, I should be able to manage it from a profitable standpoint much better. Then finally, what we always encourage people is, you know, I can start to do things on my farm because I got this report card now at the end of the year to say, you know, I was looking into this. I kind of did one field or did some strip trials like down at the bottom. And all of a sudden, you know, is that the best investment for me? Maybe I need to look at other practices. But there's a lot of potential, a lot of key uses that I can use to help me improve my operation. Talk quickly. We're not going to go in depth. Mainly today, Mark, we're going to spend time on grain yield monitors. They've been around since the early 90s. But today, there's basically two commercial available yield monitors out there. There's grain yield monitors. There's cotton yield monitoring technology. There's some other ones out there that are being developed. But I would not say they're commercially spread at this point. But cotton, we're really seeing an increase in cotton yield monitoring here in Alabama. There are essentially two ways or two companies. John Deere with their microwave technology. You can see their sensor on the bottom right. If you jump up on any of the new John Deere combines, you'll see that sensor right behind each one of the shoots. It's a plastic shoot and not a metal shoot today because of the microwave has to cannot go through the metal or it gets attenuated by the metal. The other style has been out for a while as the ag leaders and optical. Essentially, I have to cut two holes on either side of the shoot. I've got essentially a light beam, so a midder, and then I got a detector on the opposite side. And I can put those on all the shoots or a portion of the shoots. And again, all we're looking at is cotton flow through that shoot and we're estimating how much pounds per second. And we know our distance traveled. We know we're with a cut so we can calculate a pounds per acre, make an assumption on turnout and ultimately get the bales per acre in the data set. So that's cotton yield monitoring. We won't spend a lot of time. But like I mentioned, there's been a tremendous increase in cotton producers collecting yield data here in the state of Alabama in the last two years. Pretty standard today on all the John Deere combines coming out. Is it cotton pickers? Yeah, as accurate as a grain monitor would be. They do pretty good. You know, there's some things that deal with variety and the size of the bowls that you got to probably pay attention to. So calibration is required. And I would think that if you're doing a major change in terms of bowl size and density, definitely you need to recalibrate to make sure they're giving you the best results out. Not just once a season, but as you change varieties. And if you got the capacity to weigh several loads, you can essentially take a not only one basket but multiple baskets. And if you can get that weight, you can always come back and post calibrate these these cotton yield monitors, which is a nice aspect. You just got to have a way to weigh the amount that you harvest it. So we're going to jump into grain monitor grain yield monitors. Those are all the components. Again, we've got the DGPS, we've got some kind of display up in the cab. In this case, most all the grain yield monitors in the US use what we have as an impact plate at the top of the clean grain elevator. There's some other styles that use optical sensors. We'll talk a little bit about that. But then today we have all these measurements being made primarily we got grain flow, we got moisture, and we got GPS and we got ground speed or your that's those are our primary what's actually being measured. Everything else is assumed cutting with density of that grain. And we also need to know if the headers up or down that tells us if we're harvesting or not. That's just a way to start and stop data collection. So those are the components. We're going to jump down real quick mark and tell you how we calculate yield because this is an important aspect to understand. Again, thinking back what we're measuring, we're measuring mass flow, which is pounds per second. On most of these, we're measuring ground speed. So we know over a given time how far we've traveled. Okay. And so we got mass flow in the equation to calculate yield that we're measuring. We've got the test weight, which we assume and at the bottom, depending on what grain that you're doing corn, soybeans or wheat, you input that. And then we're also looking at how far we travel, you know, usually on a per second basis for over one second, if we know how far we traveled, that gives us the distance we know our swath width. And so we know the area covered so we can take that pounds per second up there on the top and calculate a bushels per acre measurement. It's simpler than it looks, john. I mean, when you when you explain it all it made sense. But what first sight that was intimidating. A little bit. But we're just making a couple measurements are cut with as an operator. You put that in I'm 25 feet on a platform I'm 20 feet, you know, or so many rows on 30 or 36 inch spacing. And so all we need is that distance traveled, we got our mass flow, and we convert that to a bushels per acre. Okay, the second thing to remember is we we have to sell this crop on grain. So we want to compute what the marketable yield. So most will have that was the wet yield. Okay, so that might be tell me what the yield is that let's say 20% moisture content. But we know when we get to the get to the market and sell, we're not going to sell it at 20%. We're going to get basically sell it for corn 15 and a half percent some cases 15%. But most of the time it's going to be 15 and a half. And so we have to we want to adjust that down to what we call the dry yield or marketable yield of most cases. And so we take that measured harvest moisture content. Okay. And you look at that equation. And basically, now I can adjust my yield bushels per acre to what I want to be able to sell. And most cases, all of our analysis are going to be done at that marketable yield or dry yields. And then we try and conform it all to 15 and a half for corn, or 13 for soybeans. So if I'm actually harvesting at 12, I'll adjust it up to 13. If I'm at 18, I just it down to 13 for soybeans. And so we're always at the same moisture content. So that's all we're doing on a on a combine, measuring a few parameters, moisture, mass flow, distance traveled, we're taking that into these equations and computing a blue bushel per acre measurement. And that's what that's what we all want to know. It's what we want to take to the coffee shop. Does the yield monitor test moisture? Or do you have I mean, we have the moisture testers that are in common use. So it's got a moisture sensor. I'll show you that here in a minute. But you got to calibrate it to typically with a high quality moisture sensor to make sure that it's given the proper moisture content to the yield calculation or the date data set. Here's what it looks like. We got a yield beta bit in most cases, all it is is point data mark. And in this case, we've chosen to kind of put a gradient green, high, red, low. But if you look up close down, all we're doing is collecting points. Typically, we're going to collect points every one second. But we can adjust that in most of these yield monitors because that's a lot of points for some of these fields. I mean, we're we're not we're not talking just 5000 we're typically you know, just a normal field 40 to 80,000 points. You think about storage and processing that. But most people keep it at one second. So every one second, I'll get data but I could adjust that to, you know, every two seconds. In some cases, there's other programs that are other displays that might give you the capacity to do it on distance. So you know, if you're if you're stopped, you know, you don't want to continue, but it'll do a distance. So every five foot, 10 foot, whatever that distance you put in, but most time it's on second, we're collecting data on a per second basis. Here's the point representation. Just really looking down even closer. And all we're saying mark, if I know the distance travel, I know my cut with, I'm taking that point and I'm assigning that area, which is essentially a rectangle. In most cases, I'm taking that area. And I'm signing that area to that single point. So just to think about that in the calculation. This is what it looks like. This is just an example, data file, raw data file that we got out of. One of the yield monitors in this case, an ag leader, you can see over the last time I got my longitude latitude, I got a mass flow measurement and pounds per second, I'm collecting that every one second. So essentially, I've got a direct, you know, way to calculate this bushels per acre. But that's what the data looks like. And there'll be other attributes field name, elevation, maybe the serial number, there's a lot of other attributes. But then a day we got our longitude, latitude, mass flow, distance traveled, typically an inches. And then we have our moisture content. And we can calculate that yield using equation for every single point. All this is done internal to the yield monitor mark. But I can get this text data out. But that's what it looks like. I just got rows and rows and rows of data ultimately, that we can all sudden process and use. That's too much data. That's why we want to keep it simple. We take a picture of the yield monitor at 120 bush on week, right? Yeah, yeah, the picture is good enough for me. So yeah. And so going back, I'm just going to go through this real quick and we'll talk. I mentioned this earlier, the impact sensor is really the primary sensing mechanism that's used in the US. But we also have these volumetric flow sensing. So there's a light sensor or photoelectric sensor that's used. We're just going to talk about those two today because those are pretty much all that we'll see in the US. The impact sensor, you'll see here, this is an example from John Deere. Grain comes up the clean grain elevator gets accelerated across there's a plate with a sensor attached to it on the top that the deflection or the momentum of that grain deflects that plate. And I basically have a relationship there between how much it deflects two pounds per second. And that's all we're doing. And again, most all these you could climb up in the grain tank, go over the top of the clean grain elevator and easily see probably a mass flow sensor sticking out. And you can take that off and you can see that that mass impact plate as we call it. Here's some examples. You see ag leader on the bottom left. And that's the most prevalent some of the older John Deere's combines. You'll see an impact plate like there in the top that's kind of yellow. Most of them are have some kind of covering on it. But there's also in the bottom right, you'll see ag codes. Again, all those are are pretty much at the top of that clean grain elevator mark. Just to mention this, most all these are not linear response. And that's why calibration will talk about calibration. But just remember, just the it's not linear. So as I press on that sensor, it doesn't give me a linear output. And so I just can't use one data point and have a calibrate. I need multiple data points to make that response curve of that particular sensor is mounted in the combine. So that's just something a lot of people forget about is this non linearity of these sensors. But we try and treat them as linear in nature. So that generates potential errors, and we don't want that. The photoelectric, you'll see this on the caterpillar combines in particular, Raven has a system, I think tremble uses very similar. But basically, you got in a light emitter that kind of about halfway up or towards the bottom of the clean grain elevator on one side, you'll see a light source. And then on the other side, directly across from it, you got basically a detector. And all you're looking as the interruption or how much interruption of light you get is green. So as we fill that paddle up, you can have more light, you know, less light transmitted. If it's empty, you got should have full light intensity in there. All we're looking at is taking that and converting that to some kind of biometric, you know, cubic meters per second. And then we take that and calculate our yield. So this is popular amongst some of the kind of third parties providing yield monitors out there. Here's your grain moisture sensor. Normally, speaking, it's going to be on the side. Here's an example of John Deere's moisture sensor. But all it's used, and most all of them use that sensor on the left there is just a DMI basic sensor. And you're measuring the capacitance or dielectric material as the moisture content goes up. Guess what, you know, you got more ions there. And so we can measure that there's a direct correlation with moisture being being able to capacitance or what we call a dielectric capability or dielectric properties change. I can measure that basically with a voltage out. And I know what the very accurately with the moisture content of that grain is. I'll take your word for John. But that's a it's this this type of sensing has been around for for some time. Does great in grain crops. And let's like most most of the companies use it. The big thing is keeping it clean. You get a little dirt little buildup on it. All sudden, you change the properties and you're going to change your measurements. Would this be like every every morning when you're blowing out your filters to wipe that down? I would definitely most of them in the it's maybe a little difficult to see there in the John Deere picture. But you'll see on the left side of kind of that black, that black box on the left side that that but basically, you there's a little wing nut, you take it off and you can look at it very quickly and wipe it down. Especially when you get into soybeans and wet soybeans or green soybeans, you will tend to get a little buildup. So every morning, you it's early season problem. You just check on it. Yeah. So so we're just going to go through some of these more just again, what we're trying to do is quality yield data. That's a real emphasis today in the in the industry is getting quality yield data. There's a lot of things. If I change my swath widths, if I if I put in my display that I'm taking a full swath with, you know, I'm taking eight rows. It thinks I'm taking eight rows, but I'm at that last row and I'm doing four. Guess what? All of a sudden, I've doubled my yield potentially. I got a time lag too. So we I think everyone recognizes when the grain hits the header to when it actually sees the impact plate or volumetric sensor, there's a timeline there. So you have to account for that. Most of the companies are doing a very good job of really getting that built in. So we know where that is. We got GPS area. That's really not a problem today. Back in the early mid 90s, we had a lot of error. And, you know, you get these maps and you couldn't tell what passes what. And so there's some other things there, pitch and roll, we'll talk a little bit about that's mainly an issue with with some of these, but not a big one. But again, companies are accounting for that, just like we're talking about the pitch and roll for GPS, we've got to if we can measure it, just like in here that we've got internal sensors, we can we can we can ensure that we're getting quality data out. And finally, calibration, I'll make a couple comments. But again, remember, you know, yield is volume per area. As simple as that. If you're taking a full header and you've told the display, you're taking a full header, then you're doing the best you can. If you're not taking a full header with guess what, all sudden that yield calculation is inflated. Yeah. Okay. And so right now, there are some things clause in particular, they can they have some sensors that monitor that. But here in the US, it's generally speaking on the operator's responsibility to tell you what the width is. If I don't change it, also new yield estimates are incorrect. So it's just something to note very quickly that that can have a huge, huge impact. Time delay. Again, as a as a user, or an operator, this isn't something necessarily ideal, that you're going to have to deal with. But most the companies but just think about it, it takes time from the header that goes through the whole heart, the threshing aspect up to clean grain elevator. Okay, and that can be sometimes eight to 11 seconds normally, it could be different for different combines. But again, most of the manufacturers are accounting for that. But the other thing to remember is the start and end delays. So when my header goes down mark, is it's starting to collect data, but you might not be already engaged in the crop. So think, you know, take you and me and take the same combine, same, same setup. I might turn around and be putting my header down at a different time than you. So just so now I can adjust those delays and we want those delays, I'll show you some examples of of errors when we talk about mapping, but we got to get those start time and end delays set. And I can adjust that based on the way I behave in that combine. train for impact style. Okay, if I'm rolling on the combine. So that's a left to right type movement. Very little influence on the impact style. This was some research that was conducted, that we reported on but basically very little influence on the measurement. But when I pitch uphill, downhill all sudden gravity starts to have an influence on on that, and definitely needs to be accounted for. Again, this really isn't an operation. An operator issue, this is what needs to be done internal to the mat to the yield monitor itself. You know, think about harvesting your yield plots, would you want to go across their up the hill? So just think about planting and what you're trying to figure out. So again, going up and down the hill, there's definitely going to be an influence on mass flow sensors. Here's just the effect of train on a photoelectric. Again, a lot of them are starting to work on this or have solutions. But think about that grain in the middle, they're going up. But also, if I get and pitch that, or start to roll that combine, excuse me, also that green slides to one side, covering up more of that light. Yeah. But yeah, I still had the same green as in as if I was flat. So that's been an issue for you for several years that most of the companies providing the photo electric type sensors have tried to build in account for. But that is a problem. Because that grain tends to get in gravity is working against you. So at the end of the day, photoelectric roll is a problem for photoelectrics on the mass flow or impact style pitch is an issue to consider. So again, if I'm going to go do and do some do some strip trials, these are just some things you might consider to make sure that you're getting data out that's best represented. The last thing we'll talk about is calibration. I already mentioned this, most of these sensors on these grain style yield monitors, or these grain yield monitors are nonlinear. So if I do one calibration, mark, we all assume linearity and look at what's happened as I as long as I'm working within that kind of normal operating conditions around that point where our calibrator is a flow of grain through that clean grain elevator, I'm doing pretty good. But if I get in the low flow, okay, a bad spot in the field, or a high spot in the field, all sudden, I want to deviate and have air in that yield measurement or mass flows measurement. Most all are doing a multi point calibration. This is pioneered by Al Meyer at Ag Leader back in the day when he developed really the first primary impact style yield monitor, where we're doing six points. And also if I can do some low flows up to high flows to kind of my normal flow rate that I expect out in the field, all sudden, no matter if I'm low or high, my data even relatively my data is going to be much more accurate. So this is why it's so important to do some low and high flow measurements in normal flow measurements as instructed in the operators manuals. Because even at a relative yield, even if we're talking about magnitude, it might not be we want that relative yield to show that the real variability out there in the field, I have to calibrate to these grain yield monitors, we're going to have errors. So in summary, you know, we just went through a lot of components today, kind of giving a kind of an overview of yield monitoring. It's the tool it's the data set that most everyone's going to want when we talk about precision ag site specific management errors do us do exist. Just be aware of that. And just think about either maybe how I plan or as I do some of my trials on farm, just just be aware of that. Okay. And today, that's what we're really driving to the focus today, and all these companies that are trying to provide services back to the farmers that are going to require yield monitor data, but the key is today quality yield data. So we're making the right decisions and helping our farmers do the right things out there. So a lot there. But one of the really important tools in the toolbox for when we talk about precision ag mark. John, this this caught my attention first yield monitors. We're just wow, we can get that data. And this is this is a exciting stuff. It's big. It's a big deal that change how you manage your farm. So thank you for this lesson. And we'll be back.